We deployed two real projects on Exovon — one built in a hurry, one built with care — and measured everything. We are sharing this because we believe the best way to learn is to see what actually happens when code meets infrastructure. Whether you are just starting out or shipping your tenth app, we hope this helps you make informed choices about your stack, your plan, and your optimization strategy.

— Ayan Safik Khan, Exovon

The Two Apps We Tested

Both are real projects with real users. One was built quickly with lots of features. The other was optimized step by step. Both teach us something useful.

	Minimanus	Deadlyvoid
What it is	A Manus-style AI agent clone	An SSR + client-side game
Code approach	Feature-first, optimize later	Built with performance in mind
Bundle size	Large (6MB+ JS)	Moderate, code-split
Images	PNG, served as-is	WebP, lazy-loaded
Dependencies	Many, some unused	Minimal, tree-shaken
Deploy region	Los Angeles (far from test users)	Pune, India (near test users)
Exovon plan	Starter (₹349)	Starter (₹349)

Both apps run on the same Exovon Pro tier. Same pricing. Same container technology. The only differences are how the code was written and where it was deployed. This tells us a lot about what you can control — and what Exovon handles for you.

Part 1: When the App Carries Extra Weight

What we learned from deploying a feature-heavy project across a long distance

Minimanus is a real, working product. It has AI agents, skill files, and a rich UI. It was built quickly to test ideas and ship features. Along the way, the bundle grew, images stayed in their original formats, and some dependencies were added but never removed. We also deployed it to Los Angeles while our test users were in India — about 13,000 kilometers away. Here is what we measured.

The Numbers

Metric	Result	What this tells us
Cold TTFB (26 min idle)	1.506 s average	The container is warm, but the round-trip is long.
Cold TTFB (worst)	2.099 s	Network variance across the Pacific adds up.
Warm TTFB (average)	1.412 s	Consistent, but still limited by distance.
Lighthouse Score (CLI)	64 / 100	Google flags this as needing improvement.
FCP	4.20 s	The browser waits for a large JS bundle to parse.
LCP	7.39 s	A big hero image and heavy layout slow things down.
Speed Index	5.90 s	The page fills in gradually rather than all at once.

These numbers are not a failure of Exovon. They are a reflection of what happens when a large bundle, unoptimized images, and a distant server all come together. We have seen similar results on other platforms with the same app. The good news: every one of these factors is within your control to improve.

If you are seeing numbers like this on your own project, do not worry. It is a common starting point. The next section shows what is possible with a few thoughtful changes.

Part 2: When Code and Geography Work Together

What thoughtful optimization and the right region can achieve

Deadlyvoid is a game. It has real-time interactions, database queries, server-side rendering, and client-side logic. But it was built with performance in mind from the start. The team used WebP for images, split the JavaScript bundle so only necessary code loads first, and removed dependencies that were not being used. It was deployed to our default serverless containers located in Mumbai (asia-south1), and tested in Pune, India — right where their players are. Here is what that looks like in practice.

The Numbers

Metric	Run 1	Run 2	Run 3	Mean	Grade
Performance Score	98	98	98	98	Excellent
FCP	1.5 s	1.5 s	1.5 s	1.5 s	Fast
LCP	2.5 s	3.4 s	2.0 s	~2.6 s	Good
Speed Index	1.5 s	2.7 s	1.5 s	~1.9 s	Fast
TBT	70 ms	120 ms	80 ms	~90 ms	Good
CLS	0	0	0	0	Perfect

A 98 Lighthouse score. Sub-two-second First Contentful Paint. Zero layout shift. This is what Exovon Pro delivers when your app is lean and deployed close to your users. It is not magic. It is the result of good habits applied consistently.

The same platform. The same container technology. The same pricing tier. The only things that changed were the code quality and the deployment region. That is empowering, because both are entirely in your hands.

Part 3: Three Lessons for Every Developer

Practical insights from running these two deployments side by side

Lesson 1: Small Code Changes Have Big Performance Impact

The difference between a 64 Lighthouse score and a 98 score is not the hosting platform. It is the bundle size, the image format, and whether you are loading only what the user needs. If your app feels slow, start here before looking at your host. Run next-bundle-analyzer. Check your image sizes. Lazy-load anything below the fold. These are free changes that often deliver the biggest speedups.

Lesson 2: Deploy Close to Your Users

Network distance matters. We hosted Minimanus in our default Mumbai (asia-south1) container pool, but we tested it from the USA. That is 13,000 kilometers of fiber, and the physical network routing adds unavoidable latency. Because all Exovon projects are currently deployed to Mumbai by default, users in South Asia will automatically experience blazing fast speeds.

FAQ: What if my major client market is in the USA?

Don't worry! While Mumbai is the current default, just send us an email and we will manually assess and migrate your workloads to a US-based cluster. We are actively building self-serve region selection directly into the Exovon UI for a future update.

Lesson 3: Choose the Right Plan for Your Stage

Exovon Free is perfect for learning, portfolios, and experiments. The container sleeps when idle, so you will see cold starts. When you are ready to scale, Starter (₹349) gives you 1GB RAM, 50GB bandwidth, and dedicated compute limits. The jump from Free to Starter is about capacity and performance for your users.

If you are ever unsure whether your app or the platform is the bottleneck, try this: deploy a simple hello-world page on the same plan and region. If that loads fast, the platform is fine and your app has room to optimize. It is a quick way to isolate the issue without guessing.

Part 4: A Practical Optimization Checklist

Five steps that will make almost any Exovon deployment faster

You do not need to do all of these at once. Pick one, measure the difference, then move to the next. Each of these changes will improve your performance on Exovon and on any other platform you might use in the future.

Step	What to do	Why it helps
1. Audit your bundle	Run next-bundle-analyzer. Look for large packages you are not using.	A smaller bundle parses faster. FCP and LCP both improve.
2. Optimize images	Convert PNG/JPEG to WebP. Use Next.js <Image> with priority on hero images. Lazy-load the rest.	Images are usually the largest assets. WebP is 25-30% smaller than JPEG.
3. Use a CDN for static files	Serve images, fonts, and CSS from Cloud CDN instead of your container.	Your container handles dynamic requests. The CDN handles static files infinitely faster.
4. Cache database queries	Use Redis for frequently accessed data. Cache SSR pages with stale-while-revalidate.	Fewer database round-trips mean lower TTFB and less backend strain.
5. Pick the right region	Deploy to the node closest to most of your users.	Shorter network paths mean lower latency. It is the easiest speed boost available.

These habits are portable. Whether you stay on Exovon for years or move to another platform tomorrow, a lean bundle, optimized images, and smart caching will always serve you well. Think of them as foundational skills, not platform-specific tricks.

Part 5: Which Exovon Plan Is Right for You?

Honest guidance based on what we have seen from real deployments

Plan	Price	Best for	What to expect
Free	₹0	Learning, portfolios, experiments	Cold starts when idle. Limited bandwidth. No database available.
Starter	₹349	MVPs, side projects, early SaaS	50GB bandwidth. 1GB RAM. Frontend Edge Hosting only (No DB Available).
Pro	₹1,499	Production SaaS, e-commerce, apps	Everything in Starter + 40GB origin bandwidth, 200 build mins & 1GB RAM. Frontend Edge Hosting only (No DB Available).
Heavy	Custom	100K+ users, heavy compute	Dedicated resources. Pay-as-you-go scaling. Direct support.

At Exovon, we are rapidly evolving and working to enhance your experience and your users' experience. We are building a platform focused on what actually matters for modern shipping: transparent pricing, powerful bundled databases, and honest performance metrics. Whether you are deploying autonomous AI agents, full-stack Next.js applications, or scalable serverless compute APIs on our global edge network, we believe developers deserve to see real numbers—not just synthetic benchmarks. If building on an AI-friendly platform that prioritizes clarity, fair pricing, and rapid deployment resonates with you, we would love to have you on board.

— Ayan Safik Khan

Founder, Exovon

August 2026

Methodology: All tests run on production Exovon infrastructure in August 2026. TTFB measured via curl. Lighthouse via Chrome DevTools (mobile) and headless CLI. Cold start = first request after 26-30 minutes idle. Warm = 5 consecutive requests. Apps are real production deployments, not synthetic benchmarks.